SummaryBACKGROUND Diagnosing tuberculosis (TB) in a human immunodeficiency virus (HIV)-endemic area is extremely difficult, as the clinical symptoms of HIV-seropositive children can be easily confounded with TB. The paediatric tuberculosis score chart (TSC) was developed for resource-poor countries and its use continues to be promoted despite the fact that this scoring system has not been evaluated in countries with a high HIV prevalence.OBJECTIVE OBJ ECTIVE To assess the utility of the TSC in an HIV-endemic area.METHOD METHO D A prospective cohort study conducted between January and December 1999 at St Theresa's Mission Hospital, Copperbelt Province, Zambia. Results of the TSC (TB score) were compared with the results of a diagnostic algorithm, incorporating sputum smear microscopy, culture and polymerase chain reaction of Mycobacterium tuberculosis, tuberculin skin test, chest X-ray and histology eventually.
RESULTSRES ULTS A total of 147 children were enrolled in the study. On the basis of HIV-serology and clinical findings they were divided into four groups: children with TB (23 HIV-seropositive; 52 HIVseronegative), 21 HIV-infected children without TB and 51 HIV-seronegative children without TB. The differences in TB scores between the groups were not significant. The sensitivity of the TSC to diagnose TB in this study was 88%; but the specificity was only 25%.CONCLUSION CONCLUSION The TSC should not be used as a diagnostic tool in countries with a high HIV prevalence. The low specificity of this scoring system leads to overdiagnosis of TB and unnecessary use of costly, antituberculous drugs. New tools for TB diagnosis in children in HIV-endemic areas are urgently needed.keywords tuberculosis, human immunodeficiency virus (HIV), children, scoring system, Zambia correspondence P. van Rheenen MD,